Full Rate

Description

Full Rate (FR) refers to the primary speech coding standard used in the GSM cellular system, officially known as the GSM Full Rate speech codec or GSM 06.10. It operates by compressing the digital speech signal to a net bit rate of 13 kbps for transmission over the radio interface. The codec is based on a Regular Pulse Excitation–Long-Term Prediction (RPE-LTP) algorithm. This algorithm models the human vocal tract to efficiently represent speech signals. The encoding process involves analyzing short frames of speech (typically 20 ms) to extract parameters like linear predictive coding (LPC) coefficients, long-term prediction (pitch) parameters, and an excitation signal represented by regular pulses.

The 13 kbps encoded bitstream is then channel coded, adding redundancy for error protection, resulting in a gross bit rate of 22.8 kbps allocated within a GSM traffic channel (TCH). This channel occupies one time slot per TDMA frame. The decoder on the receiving side reconstructs the speech signal from these transmitted parameters. The FR codec was designed to provide acceptable voice quality under a wide range of channel conditions, including moderate levels of background noise and some radio transmission errors. Its design prioritized computational simplicity for the mobile hardware of the late 1980s and early 1990s.

While FR was the baseline, GSM standards later introduced Enhanced Full Rate (EFR) and Adaptive Multi-Rate (AMR) codecs that offered superior quality or flexibility. However, FR remained a mandatory fallback option for interoperability and coverage in early GSM deployments. The term 'Full Rate' also broadly distinguishes this standard bit rate channel from the Half Rate (HR) channel, which uses a different codec to double capacity by fitting two calls into the bandwidth of one FR channel, albeit with lower speech quality.

Purpose & Motivation

The Full Rate speech codec was created as the cornerstone of GSM's voice service, the system's primary application. Its purpose was to define a digital representation of speech that could be reliably transmitted over the error-prone radio channel while meeting key constraints: limited bandwidth (one time slot per user), acceptable end-to-end voice quality, and feasible computational complexity for early digital mobile phones. Before GSM, analog cellular systems (like NMT or AMPS) did not require such a low-bit-rate digital codec.

The development of FR addressed the challenge of transitioning telephony from analog to digital within a strict spectral efficiency framework. The 13 kbps rate was a compromise between quality and capacity, enabling GSM operators to offer commercial voice services. It solved the problem of providing a standardized, interoperable voice coding method that all GSM networks and handsets could implement, ensuring seamless international roaming. While later codecs like EFR and AMR were developed to improve quality or adaptability, FR established the fundamental architecture for digital voice in mobile networks and served as the universal baseline for many years.